PostFreeSurferPipeline.sh + msm_ubuntu_3

192 views
Skip to first unread message

Christina Roßmanith

unread,
Oct 8, 2021, 5:55:05 AM10/8/21
to HCP-Users
Hi,

I'm processing my first data set using PreFreeSurfer, FreeSurfer and PostFreeSurfer. The first two of these finished successfully. But PostFreeSurfer aborts complaining that option --regoption which is set in MSMConfig is not known to msm.

I've downloaded msm from here and use HPC pipelines version Post-v4.3.0-MOD-cafd1f3.

Any hints how to solve this problem?

Kind regards,
Christina

Coalson, Timothy Scott (S&T-Student)

unread,
Oct 8, 2021, 4:54:03 PM10/8/21
to HCP-Users
Did you edit your version of SetUpHCPPipeline.sh to set this:

export MSMBINDIR=...

to where you downloaded that version of MSM?  Did you rename what you downloaded to just "msm" and chmod it to be executable?  I would expect that error from the version of MSM that ships with FSL, though it has popped up before on the list (the user built msm from a source zip, which somehow didn't match the git repo, it looks like?).

Additionally, the "MOD" in the pipelines version suggests that you have edited something, do you know what it was?  "git status" should tell you what has changed.

Tim

From: Christina Roßmanith <christina....@medma.uni-heidelberg.de>
Sent: Friday, October 8, 2021 4:55 AM
To: HCP-Users <hcp-...@humanconnectome.org>
Subject: [hcp-users] PostFreeSurferPipeline.sh + msm_ubuntu_3
 
WARNING: This message has originated from an External Source. This may be a phishing expedition that can result in unauthorized access to our IT System. Please use proper judgment and caution when opening attachments, clicking links, or responding to this email.
--
You received this message because you are subscribed to the Google Groups "HCP-Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hcp-users+...@humanconnectome.org.
To view this discussion on the web visit https://groups.google.com/a/humanconnectome.org/d/msgid/hcp-users/69ecf836-89b5-45ba-a1cd-5308c15c4227n%40humanconnectome.org.

Rossmanith, Christina

unread,
Oct 9, 2021, 8:22:32 AM10/9/21
to HCP-Users

Hi Tim,


MSMBINDIR points to HCPPIPEDIR/pipeline_tools/MSM and there msm is a link to msm_ubuntu_v3 which is executable. Even if I execute it directly it doesn't show option --regoption in its usage instructions.


The only modification of SetUpHCPPipeline.sh I've made is to set HCPPIPEDIR and MSMBINDIR (verified by git status / git diff)


What is even more strange is that I had a look at the msm github repository and couldn't find --regoption there in the source code.

Christina
Von: Coalson, Timothy Scott (S&T-Student) <tsc...@mst.edu>
Gesendet: Freitag, 8. Oktober 2021 22:53
An: HCP-Users
Betreff: Re: [hcp-users] PostFreeSurferPipeline.sh + msm_ubuntu_3
 

Glasser, Matt

unread,
Oct 9, 2021, 8:29:40 AM10/9/21
to hcp-...@humanconnectome.org

What do you get with:

 

which msm

 

Matt.

 

The materials in this message are private and may contain Protected Healthcare Information or other information of a sensitive nature. If you are not the intended recipient, be advised that any unauthorized use, disclosure, copying or the taking of any action in reliance on the contents of this information is strictly prohibited. If you have received this email in error, please immediately notify the sender via telephone or return mail.

Rossmanith, Christina

unread,
Oct 9, 2021, 5:13:28 PM10/9/21
to hcp-...@humanconnectome.org

Good news: PostFreeSurferPipeline.sh completed successfully - at least technically, I didn't check the results yet and even if 'which msm' points to the fsl version. I didn't change anything, maybe I've called the pipeline in a shell which did not source the most recent setup file. Sorry for the noise and thank you for your support.


Christina



Von: Glasser, Matt <glas...@wustl.edu>
Gesendet: Samstag, 9. Oktober 2021 14:29
An: hcp-...@humanconnectome.org

Rossmanith, Christina

unread,
Oct 11, 2021, 9:32:28 AM10/11/21
to HCP-Users

Hi,


PreFreesurferPipeline, FreesurferPipeline and PostFreesurferPipeline finished successfully for a first data set.


Now I don't understand how MEAN and MIN are calculated.

wb_command -metric-stats subj001.L.thickness.32k_fs_LR.shape.gii -reduce MAX yields 4.99 which is consistent with Matlab and R


.... -reduce MIN yields 0 but Matlab and R both report 0.3340


... -reduce MEAN prints out a value less than what Matlab and R print.


Searching through the output of 

wb_command -nifti-information subj001.thickness.32k_fs_LR.dscalar.nii -print-matrix has no match for 0.0 so I have more trust in Matlab's and R's MIN/MEAN. But what could be an explanation for the differences?


Regards,
Christina


Rossmanith, Christina

unread,
Oct 11, 2021, 9:39:31 AM10/11/21
to hcp-...@humanconnectome.org

Hi,


sorry for posting this for a 2nd time - 1st e-mail went to a wrong thread.


PreFreesurferPipeline, FreesurferPipeline and PostFreesurferPipeline finished successfully for a first data set.


Now I don't understand how MEAN and MIN are calculated by wb_command.

wb_command -metric-stats subj001.L.thickness.32k_fs_LR.shape.gii -reduce MAX yields 4.99 which is consistent with Matlab and R


.... -reduce MIN yields 0 but Matlab and R both report 0.3340


... -reduce MEAN prints out a value less than what Matlab and R print.


Searching through the output of 

wb_command -nifti-information subj001.thickness.32k_fs_LR.dscalar.nii -print-matrix has no match for 0.0 so I have more trust in Matlab's and R's MIN/MEAN. But what could be an explanation for the differences?


Regards,
Christina



Mit freundlichem Gruß,

Dipl. Inform. Christina Roßmanith
Klinik für Neurologie
Universitätsmedizin Mannheim
Medizinische Fakultät Mannheim der Universität Heidelberg
Theodor-Kutzer-Ufer 1-3
D-68167 Mannheim
Tel. +49 621 383-3551

E-Mail christina....@medma.uni-heidelberg.de
www.umm.uni-heidelberg.de

Universitätsmedizin Mannheim
universitär - modern - menschlich


Christina Roßmanith

unread,
Oct 11, 2021, 9:55:17 AM10/11/21
to HCP-Users, Christina Roßmanith
replaced metric-stats with cifti-stats and this works as expected.

Coalson, Timothy Scott (S&T-Student)

unread,
Oct 11, 2021, 4:20:24 PM10/11/21
to HCP-Users, Christina Roßmanith
To explain if it wasn't obvious, metric files don't allow excluding the medial wall, so -cifti-separate fills it with 0s.  metric files are also single-hemisphere, and don't include the subcortical data, so the results often wouldn't quite match anyway.

Tim

From: Christina Roßmanith <christina....@medma.uni-heidelberg.de>
Sent: Monday, October 11, 2021 8:55 AM
To: HCP-Users <hcp-...@humanconnectome.org>
Cc: Christina Roßmanith <Christina....@medma.uni-heidelberg.de>
Subject: [hcp-users] Re: wb_command -metric-stats
 
WARNING: This message has originated from an External Source. This may be a phishing expedition that can result in unauthorized access to our IT System. Please use proper judgment and caution when opening attachments, clicking links, or responding to this email.
--
You received this message because you are subscribed to the Google Groups "HCP-Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hcp-users+...@humanconnectome.org.

Christina Roßmanith

unread,
Oct 27, 2021, 10:34:13 AM10/27/21
to HCP-Users, tsc...@mst.edu, Christina Roßmanith
Next strange behaviour: Why does "wb_command -cifti-stats <thickness cifti> -reduce MODE" report the value 5 for data being plotted in the attached chart (visualization in wb_view)? The values for MEAN and MEDIAN make sense but MODE seems to be wrong. I'd expect something around 2.6subj001_thickness.png

Regards,
Christina

Glasser, Matt

unread,
Oct 27, 2021, 12:25:55 PM10/27/21
to hcp-...@humanconnectome.org, tsc...@mst.edu, Christina Roßmanith

Mode is the most frequent value, which could indeed be 5.000000… if there is a ceiling.

 

Matt.

 

From: Christina Roßmanith <christina....@medma.uni-heidelberg.de>


Reply-To: "hcp-...@humanconnectome.org" <hcp-...@humanconnectome.org>
Date: Wednesday, October 27, 2021 at 9:34 AM
To: HCP-Users <hcp-...@humanconnectome.org>

Cc: Timothy Coalson <tsc...@mst.edu>, Christina Roßmanith <christina....@medma.uni-heidelberg.de>
Subject: Re: [hcp-users] Re: wb_command -metric-stats

 

 

Next strange behaviour: Why does "wb_command -cifti-stats <thickness cifti> -reduce MODE" report the value 5 for data being plotted in the attached chart (visualization in wb_view)? The values for MEAN and MEDIAN make sense but MODE seems to be wrong. I'd expect something around 2.6

Harms, Michael

unread,
Oct 27, 2021, 12:29:25 PM10/27/21
to hcp-...@humanconnectome.org, tsc...@mst.edu, Christina Roßmanith

 

If that was the case, shouldn’t it show up in the histogram with the highest bar being at 5.0?

 

-- 

Michael Harms, Ph.D.

-----------------------------------------------------------

Associate Professor of Psychiatry

Washington University School of Medicine

Department of Psychiatry, Box 8134

660 South Euclid Ave.                        Tel: 314-747-6173

St. Louis, MO  63110                          Email: mha...@wustl.edu

 

From: "Glasser, Matt" <glas...@wustl.edu>
Reply-To: "hcp-...@humanconnectome.org" <hcp-...@humanconnectome.org>
Date: Wednesday, October 27, 2021 at 11:25 AM
To: "hcp-...@humanconnectome.org" <hcp-...@humanconnectome.org>
Cc: "tsc...@mst.edu" <tsc...@mst.edu>, Christina Roßmanith <christina....@medma.uni-heidelberg.de>
Subject: Re: [hcp-users] Re: wb_command -metric-stats

 

* External Email - Caution *

Glasser, Matt

unread,
Oct 27, 2021, 12:32:51 PM10/27/21
to hcp-...@humanconnectome.org, tsc...@mst.edu, Christina Roßmanith

There could be a plethora of distinct non-integer values.

 

Matt.

Coalson, Timothy Scott (S&T-Student)

unread,
Oct 27, 2021, 4:20:53 PM10/27/21
to Harms, Michael, hcp-...@humanconnectome.org, Christina Roßmanith
No, MODE doesn't do any binning, and virtually every value within the largest histogram bar is different and therefore each gets a frequency of 1 or occasionally 2 (for float32 you have 2^24 (16.8 million) possible values within a range of [2^n, 2^(n+1)], and only 32492 vertices with values (about 2^15) over a range of [1, 5], so they vastly undersample the possibilities), while all the values that were capped at 5 are the same value exactly, resulting in a frequency of probably hundreds for 5.  MODE simply doesn't make sense to use on continuous values (at least without a rounding or other binning operation first), and especially not on clamped values.  Consider if they were computed as double, with 53 bits of mantissa to work with instead of 24, the chances of an exact match would be astronomically low for any feasibly sized set of numbers produced by real-valued coordinate math.

Tim

From: Harms, Michael <mha...@wustl.edu>
Sent: Wednesday, October 27, 2021 11:29 AM
To: hcp-...@humanconnectome.org <hcp-...@humanconnectome.org>
Cc: Coalson, Timothy Scott (S&T-Student) <tsc...@mst.edu>; Christina Roßmanith <christina....@medma.uni-heidelberg.de>
Reply all
Reply to author
Forward
0 new messages